【作者】 李志伟；

【导师】 傅元略；

【作者基本信息】 厦门大学 ， 会计学， 2007， 博士

【摘要】 研究与开发(R&D)对企业的发展至关重要,但是如何对R&D项目进行评价一直以来都是一个难题。传统的现金流折现(DCF)方法由于没有考虑决策灵活性的价值,不适用于不确定性很大的R&D项目的评估。本文结合了实物期权理论与博弈论,建立了一套有效的R&D项目评价与决策的模型,并利用蒙特卡罗模拟方法,对模型进行数值求解和数据分析。本文从单企业和多企业博弈两个角度去考虑R&D项目的评价与决策。在单企业的模型中,本文充分考虑了R&D项目的特点(包括高风险性、多阶段性、可扩展性、产品生命周期特性等),并结合R&D项目进入、放弃、扩展等决策,使得所构建的模型更贴近实际。与期权理论一致,不确定性的增大将提高R&D项目的价值;在多阶段的模型中,不同阶段参数的变化对R&D项目的影响不同,对项目各个阶段的先后顺序进行合理安排将提高项目价值;项目的可扩展性将给项目带来额外的扩展期权,并大大提高项目的价值,当然项目也不应无限制的扩张,存在着一个最优的扩张规模;而当产品的生命周期特性明显时,转折点的到来会对项目的价值产生重大的影响。基于上述的研究成果,本文将模型拓展成R&D企业之间的竞争、模仿以及合作的博弈模型,企业可以根据自身的情况选择最优的博弈策略。模拟的结果证明:(1)在竞争条件下,竞争双方的价值合计小于垄断下的价值,但项目成功完成的概率大大提升;而各参数变动会对竞争对手产生反影响。(2)在模仿条件下,双方项目价值合计虽然比竞争时高,但却是以损害项目成功完成概率为代价;领先者的参数发生有利变化时,将带动模仿者的价值提升;而模仿者的参数发生有利变化,将大大损害领先者的利益。(3)在博弈策略的选择上,当双方实力相当时,双方为了争当领先者而展开激烈竞争;而当双方实力相差悬殊时,则更可能出现“领先—模仿”的均衡;当然,如果双方之间能够进行合作,只要磨合成本不是太高,合作总是可以带来更多的价值。在模型构建上,本文做了必要的改进,这些改进使得模型更加符合实际。在单企业的模型中,本文对投资成本的过程进行分解,以反映各阶段投资成本变化过程的差异;对现金流的过程进行分解,以反映生命周期各阶段需求的变化;对盈利模式的重新设计,以反映项目的可扩展性。在多企业博弈模型中,本文引入了边际成本率,以反映参与博弈企业竞争实力的差异;引入了溢出效应,以反映模仿的难易程度;引入了磨合成本,以反映合作过程中发生的额外成本。在R&D实物期权模型的求解上,本文将LSM算法(最小二乘蒙特卡罗法)与EAV算法(扩展的对偶变量法)相结合,对模型进行数值求解,解决了本文数学模型难以求出解析解的难题。这种方法相对简便,而且精确度较高,有利于模型在实际应用中推广。更多还原

【Abstract】 Research and development (R&D) is vital to the development of enterprises, but how to evaluate R&D projects has been a difficult problem. Traditional discounted cash flow (DCF) method, not considering the value of flexibility in decision-making, is unfit for the valuation of R&D projects with great uncertanty. Based on the real option theory and game theory, this dissertation develops a series of effective models of the valuation and decision of R&D projects. Subsequently, the models are solved and analyzed by the Monte Carlo method.This dissertation studies the valuation and decision of R&D projects by the view of single-firm and multi-firm. Considering some characteristics of R&D projects (including high-risk, multi-stage, expandablity, product life circle, etc.) and the decision of entry, exit and expandment, the single-firm model developed in this dissertation is more realitic. According with the option theory, the increased uncertainty will increase the value of R&D projectsIn a multi-stage model, parameters at different stages impact R&D projects in different ways, and a reasonable arrangement of various stages will enhance the value of projects. The expandablity of R&D projects will bring additional expansion options, and greatly enhance the value of projects. Of course, projects should not be expanded without limit, and there is an optimal size of the expansion. With the product life cycle characteristics, the arrival of the turning points will have a major impact on the value of projects.Based on above research results, the model is extended into the game model of competition, imitation and cooperation among firms, which can be used for choosing the optimal strategy for the game in decision making for R&D. The result of simulation shows that: (1) with the competitive conditions, the total value of both is less than the monopoly value, but the probability of successful completion of projects rises. The change of parameters has an opposite effect on competitor. (2)The total value of both under the imitative conditions is great than that under competition, which cost is the decrease of the probability of successful completion of projects. The favorable changes in the parameters of leader will help enhance the value of imitator, but that of imitator do greatly harm to the interests of leader. (3)Considering the game strategy, the two sides of equal strength will engage in fierce competition. And when strength is no match, the result may be "leading - imitation". Absolutely, if both sides could cooperate, cooperation can always bring more value with low cost.In modeling for real R&D options, there are some improvements for making these models more realistic. In the single-firm models, the process of investment cost is decomposed to reflect the difference in different stages, the process of cash flow is break up to reflect the changing needs of the various stages of the life cycle, and the mode of payoff is redesigned to reflect the expandablity of projects. In the mutli-firm game model, this dissertation introduces the marginal cost rate to reflect the strength of different firms, the spillover effect to the difficulty of imitation and the running costs to the additional cost of cooperation.The models of this dissertation, which are difficult to gain closed-form solutions, are solved by combining the LSM (Least Squares Monte Carlo) algorithm and the EAV (Extended Antithetic Variates) algorithm. This method is relatively simple with high precision, which is favor of the application of the model.更多还原